Group B Streptococcal Disease in Nonpregnant Patients: Emergence of Highly Resistant Strains of Serotype Ib in Taiwan in 2006 to 2008

Application of a perinatal disease prevention strategy based on intrapartum antibiotic prophylaxis in 1996 resulted in a substantial decline in early-onset group B Streptococcus (GBS; Streptococcus agalactiae) infections in the United States (19). The incidence of invasive GBS infections in nonpregnant patients, however, has increased over the past decades (17), particularly in elderly patients and those with underlying medical conditions (21). Diabetes mellitus, cirrhosis, and renal failure are common risk factors associated with invasive GBS disease in nonpregnant adults (11). Whereas clinical GBS isolates usually remain susceptible to penicillin, dramatic increases in resistance to erythromycin and clindamycin have raised concerns about their use as alternative agents in many countries (8, 10). The present study was undertaken to investigate the clinical characteristics of invasive GBS infection and the resistance phenotypes, capsular serotypes, and pulsed-field gel electrophoresis (PFGE) genotypes among the associated clinical isolates.

From October 2006 to June 2008, all hospitalized nonpregnant patients who were admitted to Chang Gung Memorial Hospital (CGMH) in Chiayi and had a culture-proved GBS infection were included for study. For each patient, only the first GBS isolate was subjected to laboratory investigation and analyzed. Community-acquired GBS (CA-GBS) is defined as isolates from specimens obtained within 48 h of admission, and hospital-acquired GBS (HA-GBS) as those recovered thereafter (9). The isolation of GBS from a normally sterile site suggests an invasive infection in the patient.

Double disc diffusion tests were used to determine susceptibility to erythromycin and clindamycin. Phenotypes of macrolide-lincosamide-streptogramin B (MLS_B) resistance were determined in accordance with previous reports (20). After incubation at 35.8°C for 24 h, blunting of the clindamycin inhibition zone proximal to the erythromycin disc indicated an inducible (iMLS_B) resistance. Clindamycin resistance with no blunting inhibition zones suggested a constitutive (cMLS_B) resistance. Macrolide phenotypes (M) were characterized by clindamycin susceptibility with no blunting inhibition zones around the clindamycin disc. In addition, the MICs of dalfopristin were used to identify the lincosamide-streptogramin A (LSA) phenotype in isolates with erythromycin susceptibility and clindamycin resistance. Resistance to erythromycin, clindamycin, and dalfopristin was determined by an agar dilution method according to the criteria suggested by the Clinical and Laboratory Standards Institute (7).

Capsular serotypes of the isolates were determined by a previously described multiplex PCR assay (18). SmaI (New England BioLabs, Frankfurt, Germany)-digested macrofragments of genomic DNA from serotype Ib isolates were analyzed by PFGE as previously described (25). Genotypes and subtypes were determined according to the criteria of Tenover (23).

The χ², Fisher exact χ², or Student t test was used to analyze the categorical data. Univariate and multivariate logistic regression analyses were used to discriminate independent risk factors of comorbidity. A P value of <0.05 indicates statistical significance.

A total of 228 GBS patients were included in this study (Table 1). The majority of the patients were from the age groups of 40 to 59 years (34.7%) and 60 to 79 years (35.1%). Few cases were detected in patients younger than 19 years of age (7.4%), including 11 (4.8%) infants younger than 3 months old. The female gender was predominant (64%), and most GBS diseases were noninvasive infections (all urinary tract infections) (76.8%). In contrast, invasive infections were more common in males (66.0% in males versus 26.9% in females; P < 0.01) and those over 40 years of age (26.6% among those over 40 years old compared to 11.8% among those younger).

The most common comorbidity condition was diabetes (27.6%). Multivariate analysis identified that diabetes, moderate to severe renal disease, any prior tumor, and moderate to severe liver disease were independent risk factors for invasive diseases (P < 0.01). The Charlson Comorbidity Index (CCI) is a weighted score based on the relative risk of 19 conditions significantly influencing outcomes (6). Patients were considered to have a comorbidity condition if they showed a listed disorder in the records or were treated for the disorder. The combination of the weighted scores of all comorbidity conditions present in patients was then scaled to establish the CCI. There was no significant difference in the mean CCI between genders, serotypes, and erythromycin resistance. The mean CCI of invasive diseases was significantly higher than that of noninvasive diseases (2.0 versus 0.62, respectively; P < 0.01), suggesting that the CCI is a better predictive factor for invasive GBS diseases than are individual comorbidity conditions.

The majority (94.3%) of the 228 nonrepeat GBS isolates were associated with CA-GBS, while erythromycin resistance was significantly higher in HA-GBS than in CA-GBS isolates (92.3% versus 56.3%; P < 0.01) (Table 1). Among the 12 erythromycin-resistant HA-GBS isolates, 7 showed cMLS_B resistance, and the remaining 5 showed iMLS_B resistance. Compared to CA-GBS, HA-GBS tended to cause invasive diseases (61.5% versus 20.9%; P < 0.01) and to infect elder patients (mean age, 65.1 years for HA-GBS versus 52.8 years for CA-GBS; P < 0.05). Serotype Ib was the most prevalent serotype among the HA-GBS isolates.

Serotype distribution of the 228 isolates is shown in Table 2. Resistance to erythromycin and clindamycin was found in 133 (58.3%) and 132 (57.9%) isolates, respectively. Invasiveness of the infection types was not correlated with any specific serotypes or resistance phenotypes. Compared to isolates of the other serotypes, serotype Ib isolates showed a significantly higher resistance to erythromycin (91.0% versus 47.7%; P < 0.01) and clindamycin (82.1% versus 50.0%; P < 0.01). The iMLS_B resistance type was also more frequently found in serotype Ib isolates than the others (48.1% versus 12.1%; P < 0.01).

Among the 56 serotype Ib isolates, PFGE analysis revealed six major genotypes that included a total of 17 subtypes (Fig. 1). Genotype A was the most prevalent (47 isolates [83.9%]) and contained 11 subtypes, with genotype A1 being the most prevalent (35 isolates [62.5%]). The other non-genotype A isolates were relatively more diverse, with five belonging to genotype B (2 subtypes) and one each found in genotypes C to F (Table 3). The only four isolates that were susceptible to both erythromycin and clindamycin demonstrated distinct PFGE patterns (A3, B1, D, and F). The prevalence levels of isolates that were associated with invasive diseases were similar among serotype Ib isolates of genotype A1 and those of the other PFGE patterns (17.1% versus 19.0%; P = 1.00).

Similar to previous studies from the United States (5), we observed that the majority of the GBS infections in nonpregnant patients in Taiwan occurred in adults, with an adult-to-child ratio of 12.4:1. We also found that invasive diseases were more common in male patients, usually involving subjects over 40 years of age. In the present study, risk factors for invasive GBS infection were also similar to those reported previously (11). To predict outcomes of the patients, the CCI is probably the most widely used comorbidity index to date (22), with standard clinical variables as the factors for the CCI assessment. In this study, we found that the CCI appeared useful for clinical care and risk adjustment of invasive GBS disease in nonpregnant patients.

Since the 1990s, an increase in resistance to erythromycin and clindamycin in GBS has been reported in Taiwan, from 30% and 24% in 1997 (27) to 46% and 43% in 2001 (10), and up to 58.3 and 57.9% in the present study. With a steady increase in resistance, our study also represents the first report on the significantly high incidence of erythromycin resistance (92.3% [12/13]) among HA-GBS isolates from the nonpregnant patients admitted in our hospital. Despite the small number of isolates, the results suggest that GBS is not only an established community pathogen but also could cause HA infections under high antibiotic pressure.

The MLS_B phenotype was not only prevalent among HA-GBS isolates, but it was also the most predominant resistance phenotype among the resistant GBS isolates during the study period. These data highlight the fact that target site modification by rRNA methylases, encoded by erm genes, might play a major role in the geographical and temporal variability in macrolide resistance among GBS in Taiwan. The observation is different from those reported in England, Wales, and the United States, where the macrolide phenotype is more frequent and usually involves mef genes (2, 26). Furthermore, a previous report from New Zealand indicated that the LSA phenotype was mainly found in serotype III and I isolates (14). In the present study, however, the LSA phenotype was predominant in serotype V and rarely found in serotypes Ib, II, III, IV, and VI. The presence of the LSA phenotype in different serotypes also suggested a horizontal transfer of the resistance (1).

The prevalent serotypes of GBS isolates from nonpregnant patients may vary, usually depending on the time and geographical location (3, 16). Since it was first reported in 1985, serotype V has been the most common serotype among nonpregnant adults in North America, Taiwan, and Zimbabwe (4, 12, 16). Whereas serotype Ib was an infrequent serotype elsewhere (15, 24), we found an increased isolation of serotype Ib, the second most common serotype found in this study. Our study also disclosed the high incidence of erythromycin and clindamycin resistance in isolates of this serotype (12, 13, 27). This remarkable elevation correlated with the emergence of high rates of antimicrobial resistance. The situation may be due to the clonal expansion of genotype A1 isolates in the serotype Ib population, as revealed by the conserved PFGE analysis in the present study. To our knowledge, no recent studies have examined the genetic variation of serotype Ib GBS isolates. This observation is of significant epidemiological implication. The emergence of a new genetic lineage that carries resistance to antimicrobial agents widely used in the region may represent a result of selection phenomena, as well as a therapeutic challenge. The major limitation of this study is that the isolates examined are not population based and, therefore, may not reflect the collective bias. Further surveillance of the clonal heterogeneity of GBS isolates by different measures is necessary.

• Open in new tab
• Download powerpoint

FIG. 1.

Pulsed-field gel electrophoresis patterns of serotype Ib GBS isolates. Genotype A (11 subtypes, A1 to A11) was identified in 47 isolates. Genotype B (2 subtypes, B1 and B2) was identified in five isolates. Genotypes C, D, E, and F were identified in one isolate each.

• Copyright © 2010 American Society for Microbiology